Performance Assessment of Advanced Digital Measurement and Protection Systems

Transcription

1 PSERC Performance Assessment of Advanced Digital Measurement and Protection Systems Final Project Report Part II Power Systems Engineering Research Center A National Science Foundation Industry/University Cooperative Research Center since 1996

2 Power Systems Engineering Research Center Performance Assessment of Advanced Digital Measurement and Protection Systems Final Report for PSERC Project T-22 Part II Lead Investigator: Mladen Kezunovic Graduate Students: Levi Portillo Bogdan Naodovic Texas A&M University PSERC Publication July 2006

3 Information about this project For information about this project contact: Mladen Kezunovic, Ph.D. Texas A&M University Department of Electrical Engineering College Station, TX Tel: Fax: Power Systems Engineering Research Center This is a project report from the Power Systems Engineering Research Center (PSERC). PSERC is a multi-university center conducting research on challenges facing a restructuring electric power industry and educating the next generation of power engineers. More information about PSERC can be found at the Center s website: For additional information, contact: Power Systems Engineering Research Center Arizona State University Department of Electrical Engineering Ira A. Fulton School of Engineering Phone: (480) Fax: (480) Notice Concerning Copyright Material PSERC members are given permission to copy without fee all or part of this publication for internal use if appropriate attribution is given to this document as the source material. This report is available for downloading from the PSERC website Texas A&M University. All rights reserved.

4 Acknowledgements This is Part II of the final report on the PSERC project Performance Assessment of Advanced Digital Measurement and Protection Systems (T-22). We express our appreciation for the support provided by PSERC's industrial members and by the National Science Foundation under grant NSF EEC , awarded to Texas A&M University and received under the Industry / University Cooperative Research Center program. We also thank the PSERC research project industry advisors for their technical advice. We especially thank AEP staff (Dale Krummen, John Mandeville, Dave Bernert, and Joon Park) for their support with field data, and NxtPhase staff (Dylan Stewart, Farnoosh Rahmatian and Patrick Chavez) for their support with technical information about optical sensors. i

5 Executive Summary Recently, optical voltage and current transducers, often called Optical Voltage Transformers (OVTs) and Optical Current Transformers (OCTs) respectively, have become readily available. The signals from OCTs and OVTs can be communicated to a control room through fiber optic cables. In the control room, the signals may supply a digital device, such as a relay, an energy metering system, or a power quality meter. This all-digital system may be more advantageous than conventional systems that use magnetic Current Transformers (CTs) and Potential Transformers (PTs). PTs include both Voltage Transformers and Coupling Capacitor Voltage Transformers. The optical transformers may provide improved transient response (due to a wider frequency band), improved dynamic range, and higher accuracy. This research project has sought to explore and quantify the advantages of an integrated measurement and protection system using OCTs and OVTs over a traditional system that uses conventional magnetic CTs and PTs. The comparison of the two systems was done by evaluating their relative performance when supporting the functions of protection, revenue metering, and power quality metering. This indirect approach was used, in part, because direct input/output evaluation of the optical transformers relative to magnetic transformers was not possible. The necessary input signals were not accessible in the field trials (as in most field applications) due to the high cost of retrofitting existing instrument transformers with a high accuracy referent sensor that can measure the input signals. More generally, comparison of component characteristics alone does not necessarily indicate how different component characteristics affect the performance of the functions served by those components. Hence, the approach of evaluating input/output relationships for the transformers was not pursued and instead the evaluation of the impact of different transformers on the performance of IEDs was undertaken instead. Performance indices for the functions of protection, revenue metering, and power quality metering were developed to compare the two systems quantitatively. The indirect approach was carried out using a software model of an OCT that was developed through a series of tests performed on an actual OCT in a high power lab. Coinvestigators from Arizona State University (ASU) conducted lab tests and developed the OCT model as described in a companion project report This indirect approach has a number of advantages in comparing instrument transformer characteristics based on the performance of a system that uses only the outputs from the transformer models. Simulation environment (software) can be easily expanded to evaluate new instrument transformers using their models as the transformers become available on the market. It is also possible to incorporate different power network configurations and intelligent electric devices (IEDs) The modular nature of the simulation environment makes possible the use of historical field-recorded data from different sets of instrument transformers or IEDs. ii

6 Numerical performance indices can be useful for evaluating performance of instrument transformers from different manufacturers Due to the challenges in the field data collection during this project, analysis of field data could only be used on selected cases. For the cases analyzed, the following results were obtained: Higher accuracy of optical instrument transformers with respect to conventional instrument transformers did not translate into significant improvement in the performance of the power quality and/or revenue metering IED s. Wider frequency bandwidth of the optical instrument transformers considerably improved their relative performance for relaying and metering applications. Performance of protection IED s improved when fed with signals from optical current transformers. In almost all cases, the improved performance was due to the absence of conventional CT saturation. Not all of the conventional CT models tested experienced saturation up to the level that would cause misoperation of protection IED. This suggests that the problem of saturation on protection IED performance can be avoided even when using conventional CTs by proper instrument transformer sizing. The results suggest that detailed engineering and economic analysis is required to determine the appropriateness of using an optical transformer system instead of a system with magnetic CTs and PTs. The decision to upgrade to an optical system will depend upon the performance of the conventional systems, performance of new optical transformers (in our case represented by software) models and on the objectives sought in making the upgrade. iii

7 Table of Contents 1. Introduction Literature Review and State-Of-The-Art Report Introduction Characteristics of Conventional Instrument Transformers Designs Current Transformers Voltage Transformers Accuracy Revenue Metering Accuracy Class Relaying Accuracy Class Frequency Bandwidth Current Transformers Voltage Transformers Transient Response Current Transformers Voltage Transformers Conclusion Methodology for Assessment of the Benefit of Improved Instrument Transformer Performance Characteristics Introduction Evaluation Criteria Indirect Evaluation Evaluation of Intelligent Electronic Devices Evaluation of a Measuring Algorithm Time Responses Indices Frequency Response Indices Evaluation of a Decision-Making Algorithm Use of the Criteria Evaluation Methodology Definition of Methodology Exposure Signals iv

8 3.3.3 Evaluation Based on Simulation Models Used in Simulations Power Network Current Transformer Coupling-Capacitor Voltage Transformer Overcurrent Protection Numerical Relay (Model A) Line Impedance Protection Numerical Relay (Model B) Power Quality Meter (Model C) Conclusion Assessment of the Benefit of Higher Accuracy Introduction Scenarios Evaluation Results Assessment of Higher Accuracy Conclusion Assessment of the Benefit of Wider Frequency Bandwidth Introduction Evaluation of Frequency Bandwidth Scenarios Evaluation Results Assessment of Wider Frequency Bandwidth Conclusion Assessment of the Benefit of Improved Transient Response Introduction Evaluation of Transient Response Scenarios Evaluation Results IED Model A IED Model B Assessment of Improved Transient Response IED Model A IED Model B Conclusion v

9 7. Assessment Using Field Recorded Data Introduction Sources of Field Recorded Data Evaluation of Transient Response Evaluation of Accuracy Conclusion Conclusion References Appendix Published Conference Papers vi

10 Table of Figures Figure 2.1 Mounting of HV Current Transformer... 3 Figure 2.2 Equivalent Circuit of a CCVT... 4 Figure 2.3 Mounting of HV Coupling Capacitor Volatge Transformer... 4 Figure 2.4 Frequency Response of a Current Transformer... 7 Figure 2.5 Frequency Response of a Voltage Transfomer... 7 Figure 2.6 Frequency Response of CCVT... 8 Figure 2.7 Primary Current and Electromagnetic Flux Density in the Core... 8 Figure 2.8 Primary and Secondary Currents Figure 2.9 Time-To-Saturation Curves Figure 2.10 CCVT Subsidence Transient Figure 2.11 Influence of FSC Figure 3.1 Concept of Indirect Evaluation Figure 3.2 Subfunctions of IED Figure 3.3 Flowchart of Decision Making Figure 3.4 Typical Time Response of a Measuring Algorithm Figure 3.5 Ideal and Actual Response of Measuring Algorithm Figure 3.6 Exposure Signals From a Fault Event Figure 3.7 Steps of the Simulation Approach Figure 3.8 Model of Power Network Figure 3.9 Equivalent Circuit of Current Transformer Model Figure 3.10 V-I Characteristics of Electromagnetic Core of a Current Transformer Figure 3.11 Configurations of Coupling-Capacitor Voltage Transformers Figure 4.1 Functional Elements and Flowchart of IED Model C Figure 5.1 Functional Elements and Flowchart of IED Model C Figure 6.1 Signals Associated with Abc-Phase-To-Ground Fault Figure 6.2 Comparison of Performance Index t 1max Figure 7.1 Setup for Field Data Recording vii

11 Table of Tables Table 2.1 Standard Accuracy Classes for Revenue Metering...5 Table 2.2 Secondary Terminal Voltages and Associated Standard Burdens...6 Table 3.1 Performance Indices for the Time Response of Measuring Algorithm...17 Table 3.2 Performance Indices for the Frequency Response of Measuring Algorithm...18 Table 3.3 Performance Indices - Decision Making Algorithm...19 Table 3.4 Additional Performance Indices - Decision Making Algorithm...20 Table 3.5 Parameters of Equivalent Circuit of Current Transformer Model...27 Table 3.6 Parameters of Current Transformer Models...27 Table 3.7 Parameters of Coupling-Capacitor Voltage Transformer Models...29 Table 3.8 IED and Models...29 Table 4.1 Simulation Scenario, IED Model C, Voltage Sag/Swell...33 Table 4.2 Simulation Scenario, IED Model C, Flicker...33 Table 4.3 Sag and Swell Characterization...34 Table 4.4 Flicker Characterization...34 Table 5.1 Simulation Scenario, IED Model C, Harmonics...36 Table 5.2 Simulation Scenario, IED Model C, Transients...36 Table 5.3 Harmonics Characterization...36 Table 5.4 Transients Characterization...36 Table 6.1 Simulation Scenario, IED Model A...39 Table 6.2 Simulation Scenario, IED Model B...39 Table 6.3 Current Measuring Element, ABCG Fault...42 Table 6.4 Current Measuring Element, AG Fault...42 Table 6.5 Current Measuring Element, BC Fault...42 Table 6.6 Voltage Measuring Element, ABCG Fault...43 Table 6.7 Voltage Measuring Element, AG Fault...43 Table 6.8 Voltage Measuring Element, BC Fault...43 Table 6.9 Overcurrent Decision Element, ABCG Fault...43 Table 6.10 Overcurrent Decision Element, AG Fault...44 Table 6.11 Overcurrent Decision Element, BC Fault...44 Table 6.12 Distance Decision Element, ABCG Fault...44 Table 6.13 Distance Decision Element, AG Fault...45 Table 6.14 Distance Decision Element, BC Fault...45 Table 7.1 Relay Field Recorded Data...50 Table 7.2 Power Quality Performance - 16 Recorded Sags...50 viii

12 1. Introduction This report presents results from an evaluation of instrument transformer performance characteristics based on modeling and simulation. The tasks aimed at this effort were defined in the statement of work for project titled, "Performance Assessment of Advanced Digital Measurement and Protection Systems". The tasks were defined as: Literature review and state-of-the-art report Assessment of the benefit of higher accuracy Assessment of the benefit of wider frequency band and wide dynamic range Assessment of the benefit of improved transient response on system control Assessment of operation data provided by AEP This report presents evaluation criteria, methodology and implementation of the methodology. The methodology was implemented through extensive simulation software. Simulation environment encompasses models of all the equipment involved in the evaluation. A model of the optical current transducer was developed by Arizona State University and has been included in this report. A model of the optical voltage transducer was not available. Therefore, it was not possible to evaluate performance of the optical VT model. 1

13 2. Literature Review and State-Of-The-Art Report 2.1 Introduction The first objective addresses the current state-of-the-art in the field of instrument transformers. Certain shortcomings are inherent to conventional instrument transformer designs (electromagnetic and coupling capacitor). Theoretical research and field application has shown that the mentioned shortcomings may be sufficient to cause unexpected performance of the protection, control, and monitoring subsystem applications in the electric power systems. In order to understand the shortcomings and mechanism of their influence, characteristics of conventional instrument transformer should be reviewed 2.2 Characteristics of Conventional Instrument Transformers Characteristics of the conventional instrument transformer designs (electromagnetic and coupling-capacitor) are well understood and described in the available literature [1], [2] and [3]. Operating principles are described in [1] and [2]. Historical background is given in [3]. The characteristics that define instrument transformer behavior are: Accuracy Frequency Bandwidth Transient Response Accuracy is a measure of difference between the original power network current and voltage signals and scaled-down replicas. Transient response is behavior of instrument transformers during transient power network conditions. Frequency bandwidth is a measure of maximum frequency range that can be occupied by the original power network signals to still be scaled-down correctly. The typical instrument transformer designs and characteristics are described in the sections to follow Designs Instrument transformers are available in a number of types and can be connected in a number of ways to provide the required quantities Current Transformers Current transformers are available primarily in two types: bushing and wound. Bushing transformers are usually less expensive than wound transformers, but they have lower accuracy. They are often used for relaying because of their favorable cost and because their accuracy is often adequate for relay applications. Bushing transformers are conveniently located in the bushings of power transformers and dead-tank circuit breakers, and therefore take up no appreciable space in the substation. Dead-tank circuit breakers are the preferred type of breakers in the United States, which means that they are present in much larger number than live-tank circuit breakers. Different between dead- 2

14 tank and live-tank breakers is that former ones are grounded, while the later ones are not. Because of this, live-tank breakers demand stand alone CT, i.e. CT cannot be simply mounted on the breaker; they have to be physically separated. This translates into CT for live-tank breakers being constructed in form of tall columns (consisting of insulators), isolating them from the ground. These CT are usually submerged in oil. Bushing transformers are mounted in the bushing of dead-tank breakers. They are designed with a core encircling an insulating column through which the primary current lead connects to the bushing. This means that the diameter of the core is relatively large, giving a large mean magnetic path length compared to other types. The bushing transformer also has only one primary turn, namely, the metallic connection through the center of the bushing. To compensate for the long path length and minimum primary turn condition, the cross-sectional area of iron is increased. This has the advantage for relaying that the bushing transformer tends to be more accurate than wound transformer at large multiples of secondary current rating. The bushing transformer, however, is less accurate at low currents because of its large exciting current. This makes the bushing transformer a poor choice for applications such as metering, which requires good accuracy at nominal currents Voltage Transformers Fig. 2.1: Mounting of HV Current Transformer There are two types of voltage measuring devices. They are: 1) electromagnetic voltage transformer (VT), which is a two-winding transformer, 2) coupling-capacitor voltage transformer (CCVT), which contains a capacitive voltage divider. The electromagnetic transformer is much like a conventional power transformer except that it is designed for a small constant load and hence cooling is not as important as accuracy. The coupling-capacitor device is a series stack of capacitors with the secondary tap taken from the last unit, which is called the auxiliary capacitor. The equivalent circuit of a coupling-capacitor transformer is shown in Figure 2.2 (ZB represents the transformer burden). The equivalent reactance of this circuit is defined by equation: 3

15 X L = X X C1 C1 X + X C 2 C 2 This reactance is adjusted to bring the applied voltage and the tapped voltage in phase, in which case the device is called a resonant coupling-capacitor transformer. Since the bottom capacitor is much larger that the top capacitor, i.e. X X C1 C 2 It follows that practically X L X C 2 Coupling-capacitor transformers are usually designed to reduce the transmissionlevel voltage VS to a safe metering level VB by a capacitive voltage divider, although an electromagnetic transformer may be needed to further reduce the voltage to IED voltages, usually 67 V line-to-neutral (115 V line-to-line). Fig. 2.2: Equivalent Circuit of a CCVT Fig. 2.3: Mounting of HV Coupling Capacitor Voltage Transformer 4

16 2.2.2 Accuracy There are two accuracy-rating classes for conventional instrument transformers defined in the IEEE standard [4]: Revenue Metering Class Relaying Class The definitions are based around the term transformer correction factor (TCF) [4]. TCF is the ratio of the true watts or watt-hours to the measured secondary watts or watthours, divided by the marked ratio. TCF is equal to the ratio correction factor multiplied by the phase angle correction factor for a specified primary circuit power factor. Ratio correction factor (RCF) is the ratio of the true ratio to the marked ratio. True ratio is the ratio of the root-mean-square (RMS) primary voltage or current to the RMS secondary voltage or current under specified conditions. Phase angle correction factor (PACF) is the ratio of the true power factor to the measured power factor. It is a function of both the phase angles of the instrument transformers and the power factor of the primary circuit being measured Revenue Metering Accuracy Class Accuracy classes for revenue metering are based on the requirement that the TCF of the voltage transformer or of the current transformer will be within specified limits when the power factor (lagging), of the metered load has any value from 0.6 to 1.0, under specified conditions as follows: For current transformers, at the specified standard burden at 10 percent and at 100 percent of rated primary current (also at the current corresponding to the rating factor (RF) if it is greater than 1.0). The accuracy class at a lower standard burden is not necessarily the same as at the specified standard burden. For voltage transformers, for any burden in volt-amperes from zero to the specified standard burden, at the specified standard burden power factor and at any voltage from 90 percent to 110 percent of the rated voltage. The accuracy class at a lower standard burden of different power factor is not necessarily the same as at the specified standard burden. The limits for the revenue metering accuracy classes are given in Table 2.1. Table 2.1: Standard Accuracy Classes for Revenue Metering CT VT CLASS 100% rated 10% rated Min Max Min Max Min Max

17 Relaying Accuracy Class For relaying accuracy ratings, the ratio correction will not exceed 1 percent. Relaying accuracy ratings will be designated by a classification and a secondary terminal voltage rating as follows: C, K, or T classification. C or K classification covers current transformers in which the leakage flux in the core of the transformer does not have an appreciable effect on the ratio or ratios within the limits of current and burden outlined in this item, so that the ratio can be calculated in accordance with the algebraic method (given in [4]). Current transformers with K classification will have a knee-point voltage at least 70 percent of the secondary terminal voltage rating. T classification covers current transformers in which the leakage flux in the core of the transformer has an appreciable effect on the ratio within the limits specified in item 2. An appreciable effect is defined as a 1 percent difference between the values of actual ratio correction and the ratio correction calculated in accordance with the algebraic method. Secondary terminal voltage rating. This is the voltage the transformer will deliver to a standard burden at 20 times rated secondary current without exceeding 10 percent ratio correction. Furthermore, the ratio correction will be limited to 10 percent at any current from 1 to 20 times rated secondary current at the standard burden or any lower standard burden used for secondary terminal voltage ratings. The voltage ratings and their associated burdens are as given in Table 2.2. Table 2.2: Secondary Terminal Voltages and Associated Standard Burdens SECONDARY TERMINAL VOLTAGE STANDARD BURDEN Frequency Bandwidth B-0.1 B-0.2 B-0.5 B-1 B-2 B-4 B Current Transformers Typical frequency response of a conventional CT is given in Figure 2.4 [5]. As can be seen in the figure, the transformer ratio is constants over a wide frequency range. The phase angle is also constant and has zero value. For practical purposes CT can be regarded as not having influence on the spectral content of the input signal under condition that electromagnetic flux in the core is in the linear region. In the case the flux goes out of the linear region, the change of the frequency response is hard to predict. This situation is discussed in the section 2.5. Based on frequency response of CTs, it follows that their frequency bandwidth is not limited for all practical purposes Voltage Transformers Typical frequency response of an EM voltage transformer is given in Figure 2.5 [5]. As can be seen in the figure, the transformer ratio varies significantly over wide frequency range. The phase angle also shows significant variations. Most notable sources of transformer ratio frequency dependability are: 1) stray capacitances of the primary and secondary windings, 2) stray capacitances between primary and secondary windings [6]. Typical frequency response of a coupling-capacitor voltage transformer is given in Figure 2.6 [5]. Figure also shows variations of the frequency response with the change 6

18 of various capacitances (where CC is compensating inductor stray capacitance, CP is step down transformer primary winding stray capacitance). Fig. 2.4: Frequency Response of a Current Transformer [5] Fig. 2.5: Frequency Response of a Voltage Transformer [5] Similarly as with voltage transformers, the transformer frequency response varies significantly over wider frequency range. The phase angle also shows significant variations. Most notable sources of transformer ratio frequency dependability are the same as with voltage transformers. Another factor that influences frequency response of CCVTs is the ferroresonance suppression circuit [7]. This circuit acts as a band pass filter, with center frequency at 60 Hz. More details on impact of this circuit are given in the section that deals with transient response of ITs. Frequency bandwidth of voltage VTs and CCVTs is limited. The exact limit depends on the definition of the bandwidth Transient Response The mentioned standard [4] addresses instrument transformer behavior only during the steady state and symmetrical fault power system conditions. Since behavior of instrument transformers may be significantly different for transient conditions, transient response of conventional instrument transformers has been studied. Transient response of a current transformer refers to the ability of a current transformer to handle the DC component in an asymmetrical current waveform [8]. Transient response of a voltage transformer refers 7

19 to the ability of a voltage transformer to control its tendency to create extraneous frequencies in the output [9]. Fig. 2.6: Frequency Response of CCVT [6] (a) Current Density (b) Flux Density Fig. 2.7: Primary Current and Electromagnetic Flux Density in the Core [8] 8

20 Current Transformers Saturation of the electromagnetic core is the single factor that shapes the current transformer transient response the most. Saturation may lead to signal distortions in the current transformer output. Distortion occurs whenever the core flux density enters the region of saturation. The factors influencing the core flux density are: 1) physical parameters of the current transformer, 2) magnitude, duration and waveform of the primary current signal, 3) nature of the secondary burden [8]. Saturation of the electromagnetic core can be initiated by excessive symmetrical fault currents as well as by lower magnitude asymmetrical (offset) fault currents. The fully offset fault current is shown in Figure 2.7(a). When a fully offset current is impressed on the primary of a current transformer, it will induce core flux density as shown in Figure 2.7(b) (assuming a resistive current transformer burden without loss of generality). There are two components of the total flux Φ. Alternating flux Φ ac is the flux induced by the fundamental frequency component of the fault current. Transient flux Φ tc is the flux induced by the DC component of the fault current. The variation of the transient flux Φ tc is a function of both the primary and the secondary current transformer circuit time constants. The primary current transformer circuit constant is defined by the power network section to which the current transformer is connected. The secondary current transformer circuit time constant is defined by: 1) current transformer secondary leakage impedance, 2) current transformer secondary winding impedance, 3) burden impedance. The current transformer secondary leakage impedance can usually be neglected and the current transformer secondary winding impedance is usually combined with the burden impedance to form the total burden. The dependence of the level of the saturation on the total burden is shown in Figure 2.8 [8]. The figure presents comparison between the secondary and the primary (referred to the secondary) current of a 1200:5 current transformer subjected to a fully offset current of A (20 time the rated value). In Figure 2.8(a) the current transformer is connected to the burden of Z1=(2.6+j0) Ω, while in Figure 2.8(b) the burden is Z2=(1.6+j0) Ω. It can be seen in Figure 2.8 that distortion begins certain amount of time after the fault inception. The notion of the time-to-saturation is introduced as a measure of the mentioned amount of time. The time-to-saturation is defined as the time period starting after the fault inception during which the secondary current is a faithful replica of the primary current. The time-to-saturation can be determined analytically given the power system parameters. A more practical approach is to generate a set of generalized curves that can be used for direct reading of the time-to-saturation. A set of such curves can be found in [8]. A typical set of curves is given in Figure 2.9. The set of curves is based on the current transformer primary circuit time constant T1 = 0.02 sec. Different set of curves can be obtained for a different time constant T1. The set contains curves corresponding to the current transformer secondary circuit time constant T2 ranging from 0.1 sec to 10 sec. The determination of the time-to-saturation is based on the saturation factor Ks. The factor can be calculated as: 9

21 K S VX N = I R ωt1t 2 = T T 1 2 e t T2 e t T1 + 1 Where: VX is RMS saturation voltage N2 is the number of secondary windings I1 is the primary current magnitude R2 is the resistance of total secondary burden (winding plus external resistance) ω is 2π 60 rad (a) Low Burden Voltage Transformers (b) High Burden Fig. 2.8: Primary and Secondary Currents [8] The transient response of electromagnetic voltage transformers and coupling capacitor voltage transformers depends on several distinct phenomena taking place in the primary network, such as sudden decrease of voltage at the transformer terminals due to a fault or sudden overvoltages on the sound phases during line to ground faults on the network [6]. 10

22 Sudden decrease of voltage at the primary terminals could generate internal oscillations in the windings of electromagnetic voltage transformers, which creates a high frequency on the secondary side. Fig. 2.9: Time-To-Saturation Curves [8] These high frequency oscillations are typically damped within ms. In the case of coupling-capacitor voltage transformer, energy stored in the capacitive and inductive elements of the device generate transients with low frequency of aperiodic character which could last up to 100 ms. Sudden increase of voltage at the primary terminals of electromagnetic voltage transformers could cause saturation of the magnetic core. The transient response of coupling-capacitor voltage transformers is studied in reference [9]. The study investigates the subsidence transient. The subsidence transient is the factor that influences the voltage transformer transient response the most. The subsidence transient is defined as error voltage appearing at the output terminals of a coupling-capacitor voltage transformer resulting from a sudden and significant drop in the primary voltage. The transient can be classified as belonging to one of the three classes: 1) unidirectional, 2) oscillatory, f > 60 Hz, 3) oscillatory, f < 60 Hz (see Figure 2.10). The two factors that influence the subsidence transient the most are couplingcapacitor voltage transformer burden and coupling-capacitor voltage transformer design. Fig. 2.10: CCVT Subsidence Transient [9] 11

23 Elements of the coupling-capacitor voltage transformer burden that influence the subsidence transient are: 1) burden magnitude, 2) burden power factor, 3) composition and connection of the burden. Considering the burden magnitude, most of the couplingcapacitor voltage transformer designs give smaller subsidence transient for burdens of the lower magnitude than the rated. Considering the burden power factor, the subsidence transient becomes greater as the power factor decreases, either lagging or leading. Considering the composition and connection of the burden, the following general remarks hold [9]: 1) high Q inductive elements in the burden tend to make the subsidence transient greater, 2) surge capacitors have only a minor effect on the subsidence transient, 3) series RL burdens for the same volt-ampere and power factor give smaller subsidence transient that parallel RL burdens. Fig. 2.11(a-d): Influence of FSC (a-b) resistive burden (fault initialization at zero and maximum voltage. (c-d) inductive burden (fault initialization at zero and maximum voltage) Another aspect of the transient response of coupling-capacitor voltage transformers is the impact of ferroresonance suppression circuit (FSC). This phenomenon is studied in the reference [7]. The ferroresonance is usually characterized by over voltage oscillations and distorted waveforms of current and voltage. The oscillations are mostly of subharmonic frequencies, although harmonic and even fundamental 12

24 frequencies may also be present. In order to prevent negative impact of the ferroresonance, all coupling-capacitor voltage transformers contain a ferroresonance suppression circuit, which is connected on the secondary side. FSC designs, according to their status during the transformer operation, can be divided into two main operational modes [7]: FSC in an active operation mode consists of capacitors and iron core inductors connected in parallel and tuned to the fundamental frequency. They are permanently connected on the secondary side and affect the transformer transient response. FSC in a passive operation mode consists of a resistor connected on the secondary side. This resistor can be permanently connected. Another option is to have a gap or an electronic circuit connected in series with the resistor, which are activated whenever an over voltage occurs. Such an FSC does not affect transformers transient response unless an over voltage occurs. Simulation of voltage collapse may be used as s typical example of FSC influence on the transformer transient response. The simulation results shown here are based on and FSC in active operation mode. The simulation of voltage collapse has been done using EMTP for a resistive and inductive burden of 100 Ω. Fault initiations were at the voltage zero and maximum value. The influence of the FSC is shown in Figure 2.11 (note: 1 denotes primary voltage, 2 denotes secondary voltage with FSC, referred to primary, 3 denotes secondary voltage without FSC, referred to primary). 2.3 Conclusion In this section, characteristics of typical conventional CT and VT/CCVT designs were described from the standpoint of protection system. Advantages and disadvantages of some designs over others were addressed. Three most notable instrument transformer (IT) characteristics - accuracy, frequency bandwidth and transient response, were investigated. It was shown that all three characteristics can lead to distortions in secondary waveforms that are caused by IT design characteristics. Main source of distortions with CT is saturation. Main source of distortions with VT/CCVT is subsidence transient. Causes and mechanisms of mentioned distortions were discussed. Means of lessening their impact were also addressed. The conclusion is that the impact of characteristics of the design of conventional IT on distortions is significant. When power system conditions are adverse, output signal can be significantly different from the ideal scaled-down version of input signal. 13

25 3. Methodology for Assessment of the Benefit of Improved Instrument Transformer Performance Characteristics 3.1 Introduction This section presents evaluation criteria, methodology and implementation of the methodology. Methodology was implemented through extensive simulation software. Simulation environment encompasses models of all the involved equipment. Only a model of an optical current transformer was available during creation of results for this final report. It was not possible to evaluate the optical voltage transformers. However, evaluation criteria and methodology are expandable. Models of optical voltage transformers can be readily integrated into simulation software in order to produce any results in the future. No further action is necessary beyond integration of models. To demonstrate contribution of the simulation environment to the project outcome, detailed evaluation of transient response of several conventional instrument transformers is presented, as an example of results. 3.2 Evaluation Criteria Criteria for evaluation of instrument transformers are presented in this section. First, concept of indirect evaluation is explained. Next, functional elements of power equipment are described as background for the criteria. The criteria are defined afterwards Indirect Evaluation Evaluation of the influence is done by observing behavior of control, monitoring and protection equipment when supplied with input signals coming from outputs of a particular instrument transformer under investigation. Observation of behavior means recording available output signals from the equipment and analyzing them afterwards. Objective of the analysis is extraction of performance indices. Performance indices characterize behavior of power system equipment. There are two possible approaches to evaluation: Direct evaluation Indirect evaluation Direct approach consists of comparing signals recorded on the primary side (of instrument transformers) with signals recorded on the secondary side. Primary side signals are regarded as referent signals. Since it is assumed (in this report) that signals from primary side are not available, an indirect approach for evaluation is chosen. Indirect evaluation defines criteria in the context of protection, control and monitoring functions. The concept of indirect approach is illustrated in Figure 3.1. Indirect evaluation allows comparison of behavior of a device (representing monitoring, control or protection equipment) when it is exposed to signals coming from different instrument transformers. Mentioned concept mitigates the problem of absence of referent (primary side) signals by assuming that differences in behavior of a device are 14

26 due solely to different impacts of instrument transformers. This assumption does not negate that devices can mall-function for other reasons. The assumption means that focus of this report is the influence of instrument transformer on the devices, and possible missoperations associated with the influence. UNKNOWN INPUT: KNOWN OUTPUT: Voltage signals Current signals Instrument transformer #1 Device Trip Alarm Control Data Perf. index values #1 Comparison Instrument transformer #1 Device Trip Alarm Control Data Perf. index values #2 Fig. 3.1: Concept of Indirect Evaluation Evaluation of Intelligent Electronic Devices Intelligent Electronic Devices (IED) are versatile computer-based devices employed in modern power systems for the purpose of protection, control and monitoring. Even though IED are usually designed to perform multiple functions, general subfunctions of an IED can be represented as shown in Figure 3.2 (based on reference [2]). Voltage signals Current signals Data Acquisition Measurement Decision Making Trip Alarm Control Data Fig. 3.2: Subfunctions of An IED Summary of operations of the elements shown in the figure are: "Data Acquisition" performs front-end conditioning of the input signals. Since input signals are analog current and voltage signals, Data Acquisition filters the signals using low-pass (anti-aliasing) filter, samples the signals and digitizes the signals (by converting continuous set of input values into a discrete set). In modern IEDs, data acquisition is often built as a part of the measurement element, which is explained next. "Measurement" extracts desired quantities out of input signals. Typical desired quantities are current and voltage magnitude and phase, impedance, power, direction of power flow, etc. A measuring algorithm extracts the mentioned quantities. Typical techniques used to implement the measuring algorithms are Fourier transform, Differential Equation solution, etc. 15

27 "Decision Making" derives the final output of the IED. Typical output signals are: binary (0/1) trip assertion/restrain, alarm indication, control commands, data, etc. Decision is based on a certain algorithm. The algorithm performs digital signal processing (DSP) on metered quantities, supplied by the measuring algorithm. Flowchart of Decision Making is shown in Figure 3.3. Digital processing ranges from simple comparison of values (between measured quantities and pre-set threshold values) to sophisticated artificial intelligence methods. Results of the mentioned processing are routed to the action element. Depending on the function of an IED, action element may simply output the processed data in the desirable format (e.g. in case of power measurement) or it may issue alarm or trip signal to circuit breakers (e.g. in case of protection). Threshold Quantity Metered Quantity Comparison Element/DSP Decision Element Action Element Fig. 3.3: Flowchart of Decision-Making Evaluation of IED performance is done by evaluating performance of measuring and decision making algorithms separately. Motivation for such an approach is based on design features of modern IEDs: different IEDs performing the same function can have different measuring algorithms, while different IEDs performing different functions may rely on the same measuring algorithm. Data acquisition is not evaluated separately, because it is more efficient to regard it as a part of the measurement. Evaluation of IED elements is done by recording available output signals, and analyzing them afterwards. Objective of the analysis is extraction of numerical values of output parameters. Definition of the mentioned parameters is given next Evaluation of a Measuring Algorithm Measuring algorithms extract desired parameters of the input signal. Since input to the measuring algorithm are typically current and voltage signals, typical desired parameters are magnitudes and phases of sinusoidal waveforms (based on the fundamental 50Hz/60Hz frequency). The extracted values of parameters present the response of the algorithm. The response can be evaluated in two different domains: 1) time domain, 2) frequency domain. Evaluation in the mentioned domains is discussed next Time Responses Indices Typical time response is shown in Figure 3.4. Performance indices are defined in Table 3.1. All the parameters of the indices are shown in Figure

28 Frequency Response Indices Evaluation of frequency response of a measuring algorithm involves notions of ideal and actual response. The responses are shown in Figure 3.5. Ideal response Yideal is obtained using ideal band-pass filter, to extract a set of harmonics. Actual response Yactual is very different from the ideal one. Difference is the presence of additional harmonic components in the actual response. Even though the amplitudes of additional components are typically suppressed significantly (in comparison with magnitudes of harmonic components that are being extracted), they have to be take into account when evaluation frequency response of a measuring algorithm. Based on reference [10], performance indices are defined in Table t 2% 4 2 t 1ext Fig. 3.4: Typical Time Response of a Measuring Algorithm Table 3.1: Performance Indices for the Time Response of Measuring Algorithm Index Settling time Time to the first maximum Overshoot Variable t 2 [s] t 1max [s] Δy % Definition Amount of time during which measured quantity transitions from the initial value to its steady-state value, with accuracy of 2 % Amount of time during which measured quantity reaches its first maximum value after the start of measurement y max y Δy % = y Normalized error index e norm e 1 ) L + M norm = ( y( k) y a M ( y y(0)) k = L 17

29 1.5 Y actual [p.u.] Frequency [Hz] 1.5 Y ideal [p.u.] Frequency [Hz] Fig. 3.5: Ideal and Actual Response of Measuring Algorithm Table 3.2: Performance Indices for the Frequency Response of Measuring Algorithm Index Variable Definition Gain for DC component FR DC FR = DC Y Y actual actual (0) (60) Aggregated Index F F = f 1 1 f 2 f 2 f1 Y ideal ( f ) Y actual ( f ) df Evaluation of a Decision-Making Algorithm As mentioned earlier in this chapter (see Figure 3.3), objective of a decision-making algorithm is determining status of power system, based on the measurements associated with system parameters. Changes in the power system status are associated with certain events. Event is characterized by a set of current and/or voltage waveforms, recorder (or otherwise produced) during a period of time. The period of time varies depending on IED s decision-making algorithm. Most intuitive way to evaluate performance is to evaluate how well the algorithm recognizes the power system status. Algorithm performs as expected if for a given event it recognizes a correct system state (e.g. a power quality 18

30 meter detects a certain characteristic disturbance and properly initiates necessary alarm signal). Algorithm performs unexpectedly (miss-operation), if for a specific event it makes an incorrect conclusion (e.g. a protection relay miss-interprets a minor disturbance as a fault and subsequently sends the trip command to circuit breakers). Conclusion is that decision-making criteria should reflect nature of the algorithm objective. A starting point for development (choice) of criteria is given in reference [10]. Based on the work presented in the reference, this report defines two performance indices, to serve as the criteria. Performance indices are defined in Table 3.3. Meaning of parameters in Table 3.3 is: N1 is number of events that led to correct issuance of a command signal by a decision making algorithm N0 is number of events that led to correct restrain of issuing a command signal by a decision making algorithm N is total number of events, to which IED (decision making algorithm) was exposed (during the evaluation period). A remark (discussion) is necessary about above-given parameters. In an ideal case, the equality N=N1+N0 holds. The physical meaning of the equality is: no events were unrecognized by decision-making algorithm. This is an ideal situation, which seldom occurs in actual field application of IEDs. Actual experience points to relation: N>N1+N0. Unrecognized events are detrimental to decision-making algorithm performance (i.e. the larger the value N - (N1+N0), the smaller the index S, see Table 3.4) Table 3.3: Performance Indices - Decision Making Algorithm Index Variable Definition Selectivity Average operation time S t N S = N N Amount of time between moments of event inception and decision about system status Based on the indices defined in Table 3.3, two other indices can be defined. They are defined in Table 3.4. Mentioned indices can be derived based on definitions of dependability and security, given in reference [2]. Meaning of the parameters in Table 3.4 is: N 1t is number of events for which issuing of action signal is expected N 0t is number of events for which restrain to issuing an action signal is expected

31 In ideal case, the equality N = N 1t +N 0t holds. Discussion, as the one given for parameters N1, N0, is valid also for parameters N 1t, N 0t. As can be seen, index S is function of both parameters s and d. Table 3.4: Additional Performance Indices - Decision Making Algorithm Index Variable Definition Dependability d d = N N 1 1t Security s N s = N 0 0t Use of the Criteria The criteria defined above is general in nature. Generality means that criteria defines framework for evaluation of any IED function. When evaluating influence of instrument transformers on a particular IED function, it is necessary to tailor the criteria. In the case of protective functions, the protection scheme may utilize several different zones of protection. To evaluate such function, it is necessary to evaluate protection in each of the zones. Examples of tailored criteria are presented in section 7. Similarly, in the case of power quality metering function, metering may involve detection of several different power quality events. To evaluate such function, it is necessary to evaluate detection of every event type. Use of criteria in project tasks can be summarized as following: The study addresses accuracy of optical instrument transformers. Two types of accuracies are defined in IEEE standard [11] (for more details see reference [12]): - Accuracy for protection purposes - Accuracy for metering purposes. Both types of accuracy can be evaluated using the proposed criteria for measuring algorithm. In particular, accuracy for protection purpose may be best evaluated using time response parameters, while accuracy for metering purposes can be best evaluated using both time and frequency response parameters. The study addresses frequency bandwidth and dynamic range of optical instrument transformers. Frequency bandwidth can be calculated based on frequency-response of the measuring algorithm. Dynamic range can be evaluated using time-response criteria. The two mentioned features are expected to influence mostly the power quality metering function. This function can be evaluated using the criteria for decision making algorithm. The study addresses transient response of optical instrument transformers. Transient response can be characterized by evaluating time-response of the measuring algorithm. Transient response is expected to influence mostly protection functions. This function can be evaluated using criteria for the decision making algorithm. 20